* You may need 12 minutes to read through the whole article. If you are unfamiliar with computer terms,
download the software and you'll find how easy it actually is to enable the Hardware Acceleration in the software.

Hardware acceleration^[1] hereinafter refers to a kind of efficiency promotion technology in computing field, substantially on the basis of hardware, especially GPUs (Graphics Processing Units), through execution of a set of special algorithms. More specifically, it focuses on transcoding and processing DVD videos using encoder/decoder in GPU. The article will roll out from the (positive) influence of hardware acceleration on media (video/DVD) software or say, the activity participated by such software.

Why You Need Hardware Acceleration in DVD Conversion?

DVD ripping efficiency has been increased by the leverage of CPU^[2] traditionally. Then why we choose GPU to perform acceleration during DVD conversion? What are the advantages and necessities of GPU-based hardware acceleration? The reasons mainly lie in the following aspects:

Compared with processor, the sequential structure of which makes it fail to execute some repetitive tasks in an expectedly efficient and rapid way, hardware accelerator however, is playing more roles in high-performance concurrent computing field. For example, as it is suited to intensive and massive data computing, the performances of gaming, rendering, and image algorithms have been boosted significantly thereby.

Yes, what hardware acceleration excels at is concurrent computing, for example, 192 stream processors can be operated in parallel on GeForce 200 graphics card. Coincidentally, video encoding belongs to concurrent computing. For instance, motion estimation in MPEG-2 which requires large functional block can be greatly completed by hardware acceleration depending on granularity. That's why we take the advantage.

How much faster can hardware acceleration drive for video processing? Things may be different. Take video rending with GPU^[3] acceleration as an example: the speed of rendering a 30-second AVC (.mp4) movie file with both CPU and activated GPU Processing using Sony Vegas or VideoFX Music Video Maker is 4-6 times higher than using CPU only.

What about ripping a DVD using WinX DVD Ripper Platinum while turning on the GPU accelerator? We tested ripping an encrypted 132-minute movie DVD on an average Windows OS computer.

Output Format: MP4 (H264)
Computer Condition: Windows 10 (64-bit); Intel® Core(TM) i7-8700K CPU; 16.0GB RAM
Result: The ripping process of utilizing Nvidia NVENC^[4] was completed in 8:45 minutes only, at 385 FPS averagely, as compare to 16.29 minutes at Max 201 FPS when using software-based encoding and decoding, the CPU being Intel® Core(TM) i7-8700K. Let's say, GPU acceleration can reach about 50% speed improvement compared with CPU in DVD ripping.

Disclaimer:
● This was a single DVD ripping test.
● The results may vary largely when different PC and GPU/CPU were used.

Similarly, when processing a video file that is large in size and bit rate, it will be extremely painful without hardware acceleration.

GPU encoding enables a general-purpose computer, even a low-end laptop, to convert DVDs with original image quality, because hardware acceleration has very low requirement in CPU and GPU. Currently, the absolute majority of GPUs can support the decoding of DVDs (and mainstream video formats as well) perfectly.

● For Intel: The mainstream laptops are equipped with Intel® i5 or i7 processors. That is to say, the majority of notebook users can make use of QSV for acceleration by checking "Intel" option under Hardware Accelerator item on WinX DVD Ripper Platinum (See the image, download .exe file). No matter whether your graphics cards are discrete or integrated, you can rip DVD disc to digital with hardware accelerator enabled.

● For NVIDIA: The GPU frameworks including Kepler, Maxwell (1st Gen, 2nd Gen, GM206), Pasal, Volta are supportive of NVENC-based Hardware-accelerated video encoding and decoding (See details >>). In other words, basically if you have GeForce or NVIDIA TITAN graphics card on your computer, you can make use of NVENC for faster DVD ripping. Even the old graphics card model like GeForce GTX650 is in the supported list.

See Whether Your Computer Is Hardware Acceleration Supportive >>

Intel® CPU (Intel® Ivy Bridge or higher) supports Quick Sync Video^[5] technology, which is used for video decoding/encoding. Its speed and image quality are comparable to NVIDIA CUDA or AMD Stream GPU acceleration. Also, the CPU usage is extremely low.

NVIDIA has integrated specific H.264 hardware encoder – NVENC, which is way different from the predecessor – CUDA, which realizes acceleration by executing H.264 algorithm partially though general purpose computation of GPU. NVENC, however, does encoding through the hardware unit that is customized for H.264.

You may have such experience if you are a frequent DVD user: when you play a disc, the DVD-ROM in a desktop especially in a laptop will be buzzing or heating up occasionally.

That's because during reading a disc, the optical driver kicks into high gear. In CAV mode, the reading speed of CD-ROM is roughly 200 RPM/X, of DVD-ROM 575 RPM/X (DVD reading speed is calculated by Constant Angular Velocity or Partial Constant Angular Velocity). For a DVD, 1X speed is 1.385 MB/s, equal to 1.32 MiB/s, and a 16X DVD will be 16 times higher. That explains the big noise at high (say, 10,000) RPM during DVD burning.

Different from the optical disk driver of a DVD player (hardware), the inner DVD-ROM of computer is more suitable for high-speed and continuous data reading. But if the back-end data processing is slow, the DVD-ROM will idle for most of time. Imagine how long it will idle since a DVD (DVD-9) can contain 8.5G video content! That is to say, when you play an over 2-hour movie, your DVD-ROM will be worn off for over 2 hours.

The less time DVD-ROM operates, the less time laser head will be wasted. First off, copying the DVD out for playback will already ease the burden of DVD-ROM. Additionally, with hardware acceleration, the reading time of a DVD will be greatly shortened: it only costs 5 minutes only minimally to rip a feature-length movie (the time depends also on other factors).

Furthermore, the No.1 fast and optimized DVD conversion of WinX DVD Ripper Platinum avoids idling of DVD-ROM at the extreme, reduces power consumption and prevents the temperature of DVD-ROM from being too high. Thereby, the lifespan of hardware and PC can be prolonged. According to our research, the useful life of DVD-ROM can be extended by 10 times or so. Also, your DVDs are protected for longer use.

Tips: For a longer life of DVD-ROM, you should also dissipate heat and prevent dust.

GPU-based acceleration will free the process of other apps from being influenced. The operating principle is simple: GPU serves CPU, and shares burdens of CPU. The fussy and repetitive tasks can be processed through massively parallel computing, or say, by small processors that are applied to massively parallel computing. With tasks being shunt, CPU can do more other things more efficiently without overload. Therefore, CPU can concentrate on processing other tasks. In this sense, GPU powers the acceleration of other applications.

It means if you are ripping a DVD using WinX DVD Ripper Platinum, you can work, listen to music, play games and so on concurrently, seeming like the DVD conversion runs in the background. Especially, its Level-3 Hardware Acceleration Technology will quicken and optimize the "source DVD -> HWDec -> Processing (GPU) -> HWEnc -> target video" process. See what the three levels of hardware acceleration are:

The in-operation application and computer won't crash or become slower during DVD conversion. Moreover, the temperature of CPU can be controlled in a reasonable range. During playing a DVD, the average temperature of CPU is 38 degrees centigrade. But some DVD ripper apps will raise it to as high as 80 degrees centigrade, which will make CPU out of work. But WinX DVD Ripper Platinum will let the temperature keep at 30 degrees centigrade or so averagely during conversion thanks to simultaneous CPU and GPU acceleration for decoding DVDs, and converting DVDs to MP4 H.264, DVD to HEVC/H.265, and AVI, MOV, WMV, HEVC, M2TS, MP3, iOS / Android / Windows devices, HDTV, gaming consoles, etc.

How Does DVD Ripper Software Make Use of Hardware Acceleration?

How does hardware acceleration facilitate DVD ripping process? An easy way to understand the implementation of hardware acceleration is to go through the whole decoding encoding process. Let's see the data flow chart below.

As the chart shows, it can be broken down into two big parts performed by software and hardware respectively:

- Firstly, DVD ripping software extracts video audio data block from source DVD and decrypts it into original MPEG-2 bit stream packet.
- Then it will be passed to hardware for decoding into raw data. (HWDec)
- Next, based on users' requirements of the output image, different hardware accelerators are executed to process the raw data. (Processing)
- After the processing, the data will be sent back to the hardware for encoding. (HWEnc)
- The final step: software packages the encoded video audio data into a standard multimedia file format.

Read more: Does hardware acceleration lower video quality?

No doubt that GPU assisted video encoding decoding is faster than the process solely counting on CPU. The yawning speed gap is mainly caused by architectural differences, whereas the architectures differ from each other is due to the original intentions of design. Surely, video coding is a kind of floating-point computation. But can we just jump to the conclusion that CPU is a loser when speaking of floating point arithmetic? Most of the answers lie in the architectures.

Architectures of CPU and GPU
Firstly, take a look at the next pictures of CPU/GPU architectures and see what information can be extracted.

Terminology
1. Control refers to the control unit in CPU that manages processor's all operations, including directs data, instructions and other units' operations.
2. ALU, an abbreviation of Arithmetic logic unit, is a combinational digital electronic circuit that specializes in logic and arithmetic operations.
3. Cache and DRAM (Dynamic random-access memory are components to store data.

Single Core Floating Point Computing Ability: CPU > GPU
You might be in a daze for such a result. After all, GPU, consisting of incredibly large numbers of smaller ALU cores in parallel architecture, looks every inch the more powerful data calculator than CPU. Especially, there is a saying that CPU only excels in logic operations while GPU leaves CPU in the dust in computing ability. Well, the following is the truth.

Although a CPU has only few ALU (Arithmetic Logic Unit) cores for serial operations, companied with huge control unit, its ability to perform single core floating point operation is actually far better than GPU. Each CPU core has its own cache memory for executing serial operation and is optimized for handling different tasks and doing complex equations rapidly. Even if we kick out factors like instruction set, massive cache and branch prediction, the winner still goes to CPU, for it has higher clock speed than GPU, like 4 GHz vs. 1 GHz.

Overall Floating Point Computing Ability: CPU < GPU
Yes, this is a complete reversal, and there are plausible reasons. GPU cores are good at performing same instruction in parallel at the same time but on different data, which is known as SIMD (single instruction, multiple data). Today a standard graphics card like NVidia GeForce GTX 1080 ships with a total of 2560 CUDA cores while the newest and top of the line Intel® Core™ i9-9700X X-series processor has at most 18 cores.

GPU is designed for executing multiple tasks simultaneously. It follows that, no matter how amazing a CPU core is, it can't defeat 1000 GPU cores at the same time. Of course, this only happens in specific computing tasks. But luckily, graphics processing or floating point arithmetic (for example iDCT, motion compensation and deinterlacing) is exactly one of them. And that's why hardware acceleration can be enabled while encoding decoding video.

A Much Easier Way to Understand the Math Power of CPU and GPU
Still feel dizzy? Take a look here. Suppose that CPU is a master of arithmetic, capable of addition 100 times per minute, while GPU is a schoolboy doing additions 10 times per minute. Now there are 10000 addition and subtraction that will be finished by 10 masters and 1000 schoolboys. Which team do you think will finish first? The answer is obvious. Where the GPU lacks in intelligence and complexity, it makes up in quantity. That's why hardware accelerated video encoding decoding processing is possible with GPU instead of general-purpose CPU.